Improving location of offshore earthquakes in earthquake early warning system

Hsin Chih Hsu, Da Yi Chen, Hsin Chieh Pu, Tai Lin Tseng, Yih Min Wu, Ting Li Lin

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)

Abstract

We proposed an effective approach to improve the accuracy of offshore earthquake location in the earthquake early warning (EEW) system of Taiwan. The EEW system was built upon Geiger's method for earthquake location that requires a set of initial estimates (epicenter, depth, and origin time). Because the initial epicenter highly depends on the locations of inland stations, for far offshore events the final solution falls effortlessly into a local minimum which may far away from the actual position. To solve this problem, an approach for choosing a better initial epicenter was proposed. We added predefined initial epicenters on the offshore area and then implemented several programs running Geiger's method simultaneously. Each of the programs adopted a different predefined initial epicenter. The best earthquake location is given by the most timesaving run, assuming that the solution is converged most efficiently related to the closest distance between the initial and true epicenters. The modified method has been tested with the online EEW system from June 2016 to July 2017 for offshore east Taiwan. A total of 60 earthquakes with magnitudes ranging from 3.3 to 6.0 were detected successfully. The results were compared with the estimations from the original EEW system, showing that our proposed method for offshore earthquakes is able to reduce location error by about 4.9 km on average.

Original languageEnglish
Pages (from-to)1101-1107
Number of pages7
JournalSeismological Research Letters
Volume89
Issue number3
DOIs
Publication statusPublished - 2018 May

All Science Journal Classification (ASJC) codes

  • Geophysics

Fingerprint

Dive into the research topics of 'Improving location of offshore earthquakes in earthquake early warning system'. Together they form a unique fingerprint.

Cite this